ABSTRACT
Comparisons of public opinion in England and Germany consistently show higher levels of attachment to Europe and support for European integration in Germany. The respective cross-national differences are empirically well-documented, can be framed in plausible macro-contextual explanations, and are – on the descriptive level – confirmed in the present survey of Berlin and London state secondary school head teachers. However, in view of such conclusive evidence, it is easy to overlook cross-national convergences pertaining to people's individual experiences and perceptions. This article offers an alternative view on empirically well-documented and theoretically plausible cross-national differences in levels of attachment to Europe, with a particular interest in the field of formal education. Drawing on a series of multivariate statistical analyses, it demonstrates that – in Berlin as well as in London – levels of attachment to Europe largely depend on the head teachers' intercultural experiences and skills as well as their notions of Europe as object of attachment. The respective findings lend support to educational programmes and initiatives that promote intercultural experiences and skills as means towards further European integration.
1. Attachment to Europe in England and Germany
Comparisons of public opinion in England and Germany consistently show higher levels of attachment to Europe and support for European integration in Germany (Green 2000; Carey 2002; Risse 2002; Díez Medrano 2003). And there are plausible reasons for why this ‘should’ be so.
Germany's socio-political elite has traditionally been at ease with European integration – be it for idealistic or more instrumental reasons. In the Western sphere of post-war Germany, all ‘major government and opposition parties came to view the FRG [Federal Republic of Germany] as inextricably bound within an emergent supranational community’ (Banchoff 1999: 283). Reunified Germany has maintained a strong commitment to European integration (Spohn 2002). The Western model of European integration was principally sustained and expanded to the Eastern German Länder. East Germans were quite literally ‘asked to identify with West German history’ (Ahonen 2001: 192).
In Britain (i.e., England), to the contrary, large sections of the political establishment – across different political camps – have sustained a pronounced Euroscepticism since the emergence of the European Communities after World War II to current debates over a constitutional treaty for Europe (George 2000; Garton Ash 2001; Forster 2002; Ichijo 2005; Gifford 2008). Britain's perceived loss of national sovereignty constitutes a recurrent theme in explanations of Euroscepticism in British political discourse on European integration. Kumar (2003) illustrates this point well:
For Britain, and especially for England, […] [m]embership of the European Union is perceived against a background of industrial supremacy, world empire, and victory in the Second World War. Entry into Europe therefore carries the character of a loss, if not outright humiliation, an admission that Britain is an ordinary nation, just like other nations. The implied surrender of national sovereignty is especially threatening to the English, for whom the principle of sovereignty has been a cardinal item of their national identity. (Kumar 2003: 6)
There are also important cross-national variations in media coverage on Europe and European integration (Kevin 2003). Whilst British mass media (most notably the print media) have long been predominantly Eurosceptic, (West) German mass media have traditionally been supportive of European integration (Anderson and Weymouth 1999; Díez Medrano 2003; Voltmer and Eilders 2003; McCormack 2005).
Moreover, in present-day Germany – more than in many other European countries (including Britain) – official educational narratives of the nation are anchored in supranational contexts, with a particular emphasis on processes of European integration (Soysal 2002; Soysal et al. 2005; Walkenhorst 2005). The ways in which systems of formal education construct abstract concepts such as ‘nation’ or ‘Europe’ can play an important part in the formation of (supra) national identities. After all, school curricula and textbooks ‘represent what generations of pupils will learn about their own pasts and futures as well as the histories of others’ (Soysal and Schissler 2005: 7).
2. Formal education, intercultural capital, and attachment to Europe
The link between formal education and national consciousness has long been acknowledged (Dewey 1963; Deutsch 1966; Gellner 1983; Hobsbawm 1992; Bernstein 1996). Durkheim (1961), for example, characterises the school as ‘the only moral agent through which the child is able systematically to learn to know and love his [or her] country’ (79).
More recently, visions of a more integrated Europe have been linked to expectations in the impact of schooling and school management (Pöllmann 2008b). Llobera (2001), for instance, conceives formal education as ‘one of the crucial dimensions in any attempt to develop a future European identity or at least more understanding and convergence among Europeans’ (185).
The (inter) personal benefits of intercultural teaching and learning are widely recognized and there seems to be little substantial disagreement about the most important intercultural competences – namely ‘communication and behavior in intercultural contexts’ (Deardorff 2006: 243). In most cases, ‘the more experience of other cultures a learner has, the more easily they will see the relativity of their own culture or cultures’ (Byram 2003: 65). Intercultural capital1 in form of language skills, experience of living abroad, and intercultural friendships is likely to encourage feelings of attachment beyond narrow national boundaries (Pöllmann 2009).
Teachers play an important part in shaping the views of future generations. Their individual experiences are likely to have an impact on how they appreciate and interpret curricular topics (Pajares 1992; Vulliamy et al. 1997; Pring 2001) – and ultimately on students’ learning experiences. Head teachers, as managers and key representatives of their respective school, represent a particularly interesting research population (Pöllmann 2008b).
Drawing on a survey of Berlin and London state secondary school head teachers, the present article offers an alternative view on empirically well-documented and theoretically plausible cross-national differences in levels of attachment to Europe. It suggests that – in spite of apparent cross-national variations – levels of attachment to Europe largely depend on the head teachers’ intercultural experiences and skills as well as their notions of Europe as object of attachment.
3. Survey of Berlin and London state secondary school head teachers
The data for the present study come from a postal survey on Berlin and London state secondary school head teachers’ national and European identities (Pöllmann 2008b). The concept of ‘state secondary school’ encompasses community, foundation, voluntary-aided, and voluntary-controlled schools in Greater London, along with Hauptschulen (lower secondary), Realschulen (intermediate secondary), and Gymnasien (higher secondary) in Berlin.
Between May and October 2005, the survey was administered to 624 head teachers and achieved a response rate of 45.0 percent (n =281) – 35.7 percent (n=122) in London and 56.4 percent (n =159) in Berlin. The lower response rate in London might be interpreted in terms of the different survey culture in Britain and Germany,2 the comparatively high workload of London head teachers, and the fact that Berlin head teachers received a mail package from a British university (which might have led to some courtesy responses). In both cities, however, the vast majority of respondents completed the questionnaire fully and in accordance with the format in which it was designed.
Post-fieldwork analyses of explicit refusals and response patterns over time provided no indications for topic related non-response bias. Moreover, in surveying exclusively head teachers, the likelihood of non-response bias owing to important differences in respondents’ educational background and socio-economic status appears to be negligible. And non-response bias due to inaccessibility was no cause for concern as all 624 head teachers could be contacted at their respective work address.3
4. Model and variables
This article examines levels of attachment to Europe as dependent on a series of explanatory variables. In addition to the outcome variable ‘attachment to Europe’, there are three categories of explanatory variables: the cross-national comparative dimension; four focus variables – that is explanatory variables which are of particular interest in the present article; and four control variables. Table 1 lists all relevant variables together with the respective value labels and numerical values. Appendices 1 and 3 provide information on the respective univariate distributions.
OUTCOME VARIABLE . |
---|
Attachment to Europe |
Ordinal scale: ‘not attached’ (0); ‘fairly attached’ (1); ‘very attached’ (2) |
EXPLANATORY VARIABLES |
Cross-national dimension |
City |
‘London’ (0); ‘Berlin’ (1) |
Focus variables |
Intercultural capital |
Ordinal scale: ‘low intercultural capital’ (0); ‘medium intercultural capital’ (1); ‘high intercultural capital’ (2) |
Countries considered as part of Europe |
Ordinal scale: ‘four or fewer countries’ (0); ‘five or six countries’ (1); ‘all seven countries’ (2) |
Perceived salience of Europe's cultural history |
Ordinal scale: ‘low importance’ (0); ‘medium importance’ (1); ‘high importance’ (2) |
Perceived value of EU policies |
Ordinal scale: ‘negative’ (0); ‘neither negative nor positive’ (1); ‘positive’ (2) |
Control variables |
National attachment (i.e. attachment to Britain/Germany) |
Ordinal scale: ‘not attached’ (0); ‘fairly attached’ (1); ‘very attached’ (2) |
Political alignment |
Ordinal scale: ‘far left’ (0); ‘left’ (1); ‘centre’ (2); ‘right’ (3) |
Gender |
‘Female’ (0); ‘male’ (1) |
Age |
Ordinal scale: ‘37–46’ (0); ‘47–55’ (1); ‘56–65’ (2) |
OUTCOME VARIABLE . |
---|
Attachment to Europe |
Ordinal scale: ‘not attached’ (0); ‘fairly attached’ (1); ‘very attached’ (2) |
EXPLANATORY VARIABLES |
Cross-national dimension |
City |
‘London’ (0); ‘Berlin’ (1) |
Focus variables |
Intercultural capital |
Ordinal scale: ‘low intercultural capital’ (0); ‘medium intercultural capital’ (1); ‘high intercultural capital’ (2) |
Countries considered as part of Europe |
Ordinal scale: ‘four or fewer countries’ (0); ‘five or six countries’ (1); ‘all seven countries’ (2) |
Perceived salience of Europe's cultural history |
Ordinal scale: ‘low importance’ (0); ‘medium importance’ (1); ‘high importance’ (2) |
Perceived value of EU policies |
Ordinal scale: ‘negative’ (0); ‘neither negative nor positive’ (1); ‘positive’ (2) |
Control variables |
National attachment (i.e. attachment to Britain/Germany) |
Ordinal scale: ‘not attached’ (0); ‘fairly attached’ (1); ‘very attached’ (2) |
Political alignment |
Ordinal scale: ‘far left’ (0); ‘left’ (1); ‘centre’ (2); ‘right’ (3) |
Gender |
‘Female’ (0); ‘male’ (1) |
Age |
Ordinal scale: ‘37–46’ (0); ‘47–55’ (1); ‘56–65’ (2) |
The outcome variable ‘attachment to Europe’ consists of three ordinal categories. Changes on this outcome variable will be examined by estimating an ordered logistic regression model, followed by a number of post-estimation procedures.4
4.1. Attachment to Europe
A four-category ordinal scale was used to ‘measure’ the head teachers’ levels of attachment to Europe (i.e., ‘not at all attached’; ‘not very attached’; ‘fairly attached’; and ‘very attached’).5 Due to the small number of empirical observations in the original category ‘not at all attached to Europe’ (three cases in Berlin; four cases in London), the original categories ‘not very attached’ and ‘not at all attached’ were merged into a new category ‘not attached’ as outlined in Table 1.
Table 2 shows marked cross-national differences in levels of attachment to Europe. In particular, the percentage of London head teachers who do not feel attached to Europe (30.3 percent) is about twice as high as the corresponding percentage for Berlin head teachers (15.7 percent). But this is far from unexpected. These results merely confirm empirically well-documented and theoretically plausible cross-national differences in levels of attachment to Europe.
. | Berlin . | London . |
---|---|---|
Very attached | 28.3% | 23.0% |
Fairly attached | 56.0% | 46.7% |
Not attached | 15.7% | 30.3% |
Valid observations | 159 | 122 |
. | Berlin . | London . |
---|---|---|
Very attached | 28.3% | 23.0% |
Fairly attached | 56.0% | 46.7% |
Not attached | 15.7% | 30.3% |
Valid observations | 159 | 122 |
However, it is worth looking beyond the surface of univariate comparisons and macro-level explanations. As the subsequent multivariate analyses will demonstrate, perceptions of Europe as object of attachment and levels of intercultural capital have a significant impact on levels of attachment to Europe – irrespective of the head teachers’ national background.
4.2. Intercultural capital
To generate the intercultural capital variable, a sum score was computed for each head teacher based on the three components of intercultural capital – i.e., language skills, experience of living abroad, and intercultural friendships (see Appendix 2). To achieve an appropriate ratio of empirical cases to categories of explanatory variables, the number of levels of intercultural capital was subsequently reduced from eight to three which lead to the final operational measure, as outlined in Table 1 and Appendix 1.6
4.3. Notions of the cultural, geopolitical, and institutional Europe
European integration after World War II may be associated with economic development, mobility, and peace (Dinan 2005; Giddens 2006; Pöllmann 2007). But the vision of a more culturally and socially integrated Europe may also be characterised as ‘a dream confined to some intellectuals’ (Smith 1995: 24) – an elusive fiction of an elitist circle of Euro-bureaucrats. Different people are likely to imagine different things when thinking of ‘Europe’. Geopolitical ideas of Europe, for example, do not necessarily converge with the borders of the European Union or the ‘European continent’ as delineated in atlases (Delanty 1995). The analyses will therefore be expanded beyond explorations of the extent to which the head teachers feel attached to Europe.
The inherent relationality of ‘attachment’ (i.e., ‘to feel attached to x’) allows one to analyse levels of attachment in relation to a specific object of attachment (i.e., Europe).7 Three factors can be considered that shall account for (some of the) variation in the respondents’ notions of Europe – namely Europe's perceived geopolitical inclusiveness, the perceived salience of Europe's cultural history, and the value attributed to different EU policies.
To tap into subjective notions of Europe's geopolitical inclusiveness, the head teachers were asked to specify whether (or not) they perceive Russia, Britain, Estonia, Iceland, Turkey, the Czech Republic, and Germany as part of Europe. The selection of countries was partly inspired by the European Identity Survey (2003) and shall represent Europe's cultural and geopolitical diversity. Based on this primary data, the total sum of countries named as part of Europe was computed for each individual respondent. This led to an intermediate measure with a range from ‘0’ (no country included) to ‘7’ (all seven countries included). In a further transformation step, this intermediate measure was recoded into a more workable three-point ordinal scale, as listed in Table 1 and Appendix 1.8
In addition, associations between the head teachers’ levels of attachment to Europe and the extent to which they perceive Europe's cultural history as important or not will be examined. Table 1 provides information on the ‘cultural history variable’ together with value labels and numerical values. Appendix 1 lists the percentages for the different levels of importance.9
To take subjective notions of the institutional Europe into account, the head teachers were asked to evaluate a selection of different EU policy areas – namely: protection of the global environment; growth of the national economy; the fight against poverty in the world; national social welfare; the fight against international terrorism; the protection of ethnic minority groups in Britain/Germany; the promotion of human rights; and the rights of non-EU citizens with residence in Britain/Germany. For each of these eight policy areas, the head teachers could indicate whether they perceived the European Union's role as negative, neither negative nor positive, or as positive. For each head teacher the sum of her10 evaluations of the eight initial empirical indicators was computed. This resulted in an ‘intermediate measure’ with 17 categories. The number of categories was subsequently reduced from 17 to three, as outlined in Table 1 and Appendix 1.11
4.4. Control variables
National attachment12 constitutes an important control variable in the statistical model. Whilst empirical research in Germany largely suggests the reconcilability of feelings of attachment to country and feelings of attachment to Europe (Florack and Piontkowski 2000; Pöllmann 2007, 2008a), it is rather atypical to find positive associations between national and supranational forms of attachment in British, especially English, samples (Duchesne and Frognier 1995; Cinnirella 1997; Huici et al. 1997; Carey 2002). It will be interesting to see whether the findings for Berlin and London head teachers reflect these wider national trends.
In addition, political alignment, gender, and age will be considered – all of which represent standard variables in research on (supra) national forms of attachment or identity. In terms of people's position on European integration, for example, it has been shown that ‘left identification grew stronger as a predictor of positive attitudes toward the EU while right identification became an even more powerful predictor of negative attitudes’ (Nelsen et al. 2001: 200). The findings for gender and age are less conclusive (Deflem and Pampel 1996; Nelsen and Guth 2000; Jones and Smith 2001; Carey 2002).
Other widely used control variables include educational background, income, and degree of urbanity.13 Moreover, comparisons across different time periods have demonstrated significant variation in levels of (supra) national attachment over the years (Green 2000; Nelsen et al. 2001; Citrin and Sides 2004). However, given that all respondents are state secondary school head teachers, their educational background and income (though certainly not identical) can be conceived as quasi constants. Further, quasi-constant degrees of urbanity in Berlin and Greater London were assumed, and the cross-sectional nature of the survey data precludes comparisons across time. 14
5. Data analysis
Table 3 contains the ordered logit coefficients for each explanatory variable in the model, the associated levels of statistical significance, standard errors of the respective logit coefficients, two cut points, the log likelihood (with intercept only and for the full model), the χ2 for the overall model, Brant test statistics, the number of valid observations, as well as the number of missing cases.
. | Coefficients . | SE . |
---|---|---|
Cross-national dimension | ||
City | 1.207*** | 0.287 |
Focus variables | ||
Intercultural capital | 0.686*** | 0.174 |
Countries considered as part of Europe | 0.155 | 0.207 |
Perceived salience of Europe's cultural history | 0.892*** | 0.224 |
Perceived value of EU policies | 0.558** | 0.194 |
Control variables | ||
National attachment | 1.628*** | 0.232 |
Political alignment | –0.631*** | 0.167 |
Gender | –0.113 | 0.266 |
Age | 0.057 | 0.188 |
Cut point 1 (SE) | 3.284 (0.717) | |
Cut point 2 (SE) | 6.350 (0.802) | |
Log likelihood with intercept only | –280.856 | |
Log likelihood of full model | –228.826 | |
LR chi2 (9) | 104.06*** | |
Brant chi2 (9) | 12.40a | |
Number of observations | 274 | |
Number of missing cases | 7 |
. | Coefficients . | SE . |
---|---|---|
Cross-national dimension | ||
City | 1.207*** | 0.287 |
Focus variables | ||
Intercultural capital | 0.686*** | 0.174 |
Countries considered as part of Europe | 0.155 | 0.207 |
Perceived salience of Europe's cultural history | 0.892*** | 0.224 |
Perceived value of EU policies | 0.558** | 0.194 |
Control variables | ||
National attachment | 1.628*** | 0.232 |
Political alignment | –0.631*** | 0.167 |
Gender | –0.113 | 0.266 |
Age | 0.057 | 0.188 |
Cut point 1 (SE) | 3.284 (0.717) | |
Cut point 2 (SE) | 6.350 (0.802) | |
Log likelihood with intercept only | –280.856 | |
Log likelihood of full model | –228.826 | |
LR chi2 (9) | 104.06*** | |
Brant chi2 (9) | 12.40a | |
Number of observations | 274 | |
Number of missing cases | 7 |
Notes: **P<0.01; ***P<0.001.
aGlobal and individual Brant test statistics are not significant, which indicates that the parallel regression assumption has not been violated.
The χ2 value for the overall model is 104.06, which is statistically significant at less than the 0.001 level and indicates a good model fit. The Brant test statistic is not significant on the 0.05 level, which means that the proportional odds assumption has not been violated.15
The coefficients for three out of four focus variables are statistically significant – intercultural capital (p <0.001), the perceived salience of Europe's cultural history (p <0.001), and the perceived value of EU policies (p <0.01). In contrast, the number of countries that the head teachers perceive as part of Europe does not have a statistically significant impact on the extent to which they feel attached to Europe.
National attachment and political alignment emerge as statistically significant control variables (both with p <0.001). The further to the ‘political right’ a respondent perceives herself and the less she feels attached to Britain/Germany, the more likely she is to not feel attached to Europe. The head teachers’ gender and age, on the other hand, do not have a statistically significant impact on their levels of attachment to Europe.
In what follows, predicted probabilities will be used to better illustrate the explanatory impact of the three statistically significant focus variables. Predicted probabilities are relative to the specific configuration of all explanatory variables in the model. In other words, ‘the magnitude of the change in the outcome probability for a given change in one of the independent [explanatory] variables depends on the levels of all of the independent [explanatory] variables’ (Long and Freese 2003: 151). The predicted probabilities of a given outcome can vary between values of ‘0’ (outcome is perfectly unlikely) and ‘1’ (outcome is perfectly likely).
5.1. Cross-national comparisons of changes in predicted probabilities
At this stage, primary interest lies in those respondents who ‘lean’ towards either end of the three-point attachment scale – that is, those head teachers who feel either very or not attached to Europe. The outcome ‘fairly attached to Europe’ will feature in subsequent analyses.
The line charts in each figure illustrate how the predicted probabilities (on y-axes) of the respective outcome change according to different values of the respective focus variable (on x-axes), holding all other explanatory variables at their means. For purposes of illustration, these lines connect discrete values of the respective focus variable – they do not constitute linear relationships. In each chart, the line with disks represents the results for Berlin head teachers, and the line with diamonds the results for London head teachers. Due to constraints in terms of space, the value labels could not be included in the respective figures; however, Table 1 does provide this information.
As Figure 1 shows, higher levels of intercultural capital markedly raise the predicted probabilities of the outcome ‘very attached to Europe’, whilst reducing the likelihood of a respondent feeling not attached to Europe. Moreover, as Figure 2 illustrates, the more salience the head teachers ascribe to Europe's cultural history and the more positively they evaluate various EU policies, the more likely they are to feel very rather than not attached to Europe. In all instances, the trends for Berlin and London head teachers are largely comparable.
Intercultural capital as predictor of attachment to Europe
Note: ‘Pr’ stands for predicted probabilities (on y-axes); Table 1 contains descriptive labels for numerical values (on x-axes).
Intercultural capital as predictor of attachment to Europe
Note: ‘Pr’ stands for predicted probabilities (on y-axes); Table 1 contains descriptive labels for numerical values (on x-axes).
Notions of Europe as predictor of attachment to Europe
Note: ‘Pr’ stands for predicted probabilities (on y-axes); Table 1 contains descriptive labels for numerical values (on x-axes).
Notions of Europe as predictor of attachment to Europe
Note: ‘Pr’ stands for predicted probabilities (on y-axes); Table 1 contains descriptive labels for numerical values (on x-axes).
5.2. Three ideal types
To demonstrate the joint explanatory impact of intercultural capital, the perceived salience of Europe's cultural history, and the perceived value of EU policies, three ideal typical configurations of explanatory variables will now be considered. These ideal types serve as analytical tools in order to illustrate how different configurations of explanatory variables alter the predicted probabilities of different levels of attachment to Europe. Whilst they do constitute abstract configurational thought experiments, the respective predicted probabilities are grounded in real empirical data.16Table 4 lists the respective predicted probabilities by city, considering all three levels of attachment.
Ideal type 1: . | ||
---|---|---|
All explanatory variables at average values | ||
Berlin | London | |
Predicted probability of ‘very attached to Europe’ | 0.29 | 0.11 |
Predicted probability of ‘fairly attached to Europe’ | 0.61 | 0.62 |
Predicted probability of ‘not attached to Europe’ | 0.10 | 0.28 |
Ideal type 2: | ||
Intercultural capital (high intercultural capital); perceived salience of Europe's cultural history (high importance); perceived value of EU policies (positive); rest (average values) | ||
Berlin | London | |
Predicted probability of ‘very attached to Europe’ | 0.62 | 0.33 |
Predicted probability of ‘fairly attached to Europe’ | 0.35 | 0.59 |
Predicted probability of ‘not attached to Europe’ | 0.03 | 0.09 |
Ideal type 3: | ||
Intercultural capital (low intercultural capital); perceived salience of Europe's cultural history (low importance); perceived value of EU policies (negative); rest (average values) | ||
Berlin | London | |
Predicted probability of ‘very attached to Europe’ | 0.02 | 0.01 |
Predicted probability of ‘fairly attached to Europe’ | 0.30 | 0.12 |
Predicted probability of ‘not attached to Europe’ | 0.67 | 0.87 |
Ideal type 1: . | ||
---|---|---|
All explanatory variables at average values | ||
Berlin | London | |
Predicted probability of ‘very attached to Europe’ | 0.29 | 0.11 |
Predicted probability of ‘fairly attached to Europe’ | 0.61 | 0.62 |
Predicted probability of ‘not attached to Europe’ | 0.10 | 0.28 |
Ideal type 2: | ||
Intercultural capital (high intercultural capital); perceived salience of Europe's cultural history (high importance); perceived value of EU policies (positive); rest (average values) | ||
Berlin | London | |
Predicted probability of ‘very attached to Europe’ | 0.62 | 0.33 |
Predicted probability of ‘fairly attached to Europe’ | 0.35 | 0.59 |
Predicted probability of ‘not attached to Europe’ | 0.03 | 0.09 |
Ideal type 3: | ||
Intercultural capital (low intercultural capital); perceived salience of Europe's cultural history (low importance); perceived value of EU policies (negative); rest (average values) | ||
Berlin | London | |
Predicted probability of ‘very attached to Europe’ | 0.02 | 0.01 |
Predicted probability of ‘fairly attached to Europe’ | 0.30 | 0.12 |
Predicted probability of ‘not attached to Europe’ | 0.67 | 0.87 |
The first ideal type shall demonstrate the effect of neutralising the explanatory impact of all statistically significant explanatory variables (i.e., controlling for all explanatory variables). As a result, ideal type 1 mirrors univariate cross-national divergences in overall levels of attachment to Europe (see Table 2).
Ideal type 2 represents a respondent for whom Europe's cultural history is important in terms of what she means when thinking of Europe. Furthermore, she perceives EU policies as positive and shows high levels of intercultural capital – with average values on all other explanatory variables. As Table 4 illustrates, this configuration of explanatory variables manifestly changes the predicted probabilities – in both cities – in favour of the outcome category ‘very attached to Europe’. In Berlin as well as in London, the predicted probabilities for the outcome category ‘not attached to Europe’ are now close to zero.
As the inverse of ideal type 2, the third ideal type stands for a respondent who perceives Europe's cultural history as unimportant in terms of what she means when thinking of Europe. Additionally, she regards EU policies as negative and shows low levels of intercultural capital – with average values on all other explanatory variables. Table 4 illustrates that this configuration of explanatory variables significantly alters the respective predicted probabilities. In Berlin as well as in London, the category ‘not attached to Europe’ now constitutes the by far most likely outcome. On the other hand, the predicted probabilities for the outcome category ‘very attached to Europe’ approximate zero in both cities.
6. Conclusions
Cross-national differences between levels of attachment to Europe in England and levels of attachment to Europe in Germany are empirically well-documented, can be framed in plausible explanations, and are – on the descriptive level – confirmed by the present survey of Berlin and London state secondary school head teachers (see Table 2). But despite apparent cross-national differences, Berlin and London head teachers’ attachments to Europe are comparable on a number of accounts.
As Appendix 1 illustrates, some respondents have less inclusive notions of the geopolitical Europe than others. But in both cities there is no statistically significant association between the number of countries that the head teachers perceive as part of Europe and the extent to which they feel attached to Europe (see Table 3). This implies that high levels of attachment to Europe do not need to coincide with narrow geopolitical definitions of (a fortress) Europe.
In both cities, levels of attachment to Europe fluctuate depending on how the head teachers value Europe's cultural history and a number of different EU policies (see Figure 2). Irrespective of national background, the more positively a head teacher sees the respective EU policy areas and the more salience she attributes to Europe's cultural history,17 the more likely she is to show high levels of attachment to Europe.
Cross-national similarities also involve the relationship between levels of national and supranational attachment. In both cities, high levels of national attachment do not translate into low levels of attachment to Europe, or vice versa (Pöllmann 2008a). In fact, levels of national attachment and levels of attachment to Europe show a strong positive association (see Table 3), suggesting that one form of attachment may further, rather than obstruct, the appropriation of another (Guetzkow 1955; Delanty and Rumford 2005).
Intercultural capital in form of language skills, experience of living abroad, and intercultural friendships has a significant positive impact on levels of attachment to Europe. Certainly, levels of intercultural capital vary in terms of both degree (see Appendix 1) and content (see Appendix 2). The average Berlin head teacher speaks more foreign languages than her London colleague. London head teachers, on the other hand, report on average a more intercultural circle of friends as well as more living experience in other countries. But irrespective of national background, higher levels of intercultural capital are associated with higher levels of attachment to Europe. Lower levels of intercultural capital have a reverse explanatory impact (see Figure 1).
The analysis of three distinctly different ideal types served to illustrate the joint explanatory impact of intercultural capital and notions of Europe as object of attachment. In such a constellation, otherwise apparent cross-national differences in the predicted probabilities for different levels of supranational attachment shrink significantly – particularly for the outcomes very attached and not attached to Europe (see Table 4). In the end, apparent cross-national differences in levels of attachment to Europe turn out to be less clear-cut than they may appear at first glance.
These findings have implications for the field of schooling and school management. The explanatory impact of intercultural capital lends support to educational programmes and initiatives that promote intercultural experiences and skills (e.g., through international student and staff exchanges) as means towards further European integration. Moreover, a better understanding of the relationship between people's notions of the cultural, institutional, and geopolitical ‘Europe’ – and the extent to which they feel attached to Europe – could inform the teaching and learning of ‘European topics’ in various subject areas including citizenship education, history, geography, sociology, and language education.
Of course, the cross-national comparisons presented in this article are relative to a specific conceptual approach; to specific methods of data collection and analysis; and to specific samples of head teachers in two specific cities at a specific point in time. Further comparative research on predictors of people's attachment to Europe will have to show whether the explanatory value of intercultural capital and notions of Europe as object of attachment can be generalised to different people, places, and times.
Footnotes
See Pöllmann (2009) for a more detailed discussion of ‘intercultural capital’.
There is a general trend towards higher response rates in surveys conducted in Germany. Furthermore, recent research suggests that response rates are particularly low from London schools (Lynn and Egerton 2002).
For more information on concepts, operational measures, choice of research populations, sampling, data collection, response rates, and variable transformations, see Pöllmann (2008b).
I computed a number of post-estimation procedures known as ‘SPost’ (for Stata Post-estimation commands) and written by J. Scott Long and Jeremy Freese. SPost is available for download from within Stata® (i.e., ‘net search spost’). For further details, see Long and Freese (2003).
The recoding procedure reads as follows: ‘0–2’ into ‘0’, ‘3–4’ into ‘1’, and '5–7’ into ‘2’. See Pöllmann (2008b) for information on the respective survey questions and a discussion of univariate results.
‘Identification’ (other than ‘identity’) would have done a similar job since it logically implies an object of identification (i.e., ‘to identify with x’).
The recoding procedure reads as follows: ‘0–4’ into ‘0’, ‘5–6’ into ‘1’, and '7’ into ‘2’. See Pöllmann ( 2008b) for information on the respective survey question and a discussion of univariate results.
To achieve an appropriate ratio of empirical cases to categories of explanatory variables, a new three-point ordinal scale was created, as outlined in Table 1 and Appendix 1. The recoding procedure reads as follows: ‘0-1’ into ‘0’, ‘2’ into ‘1’, and '3-4’ into ‘2’. See Pöllmann (2008b) for information on the respective survey question and a discussion of univariate results.
When mentioned in relation to head teachers, ‘her’ stands for ‘his or her’ – and ‘she’ for ‘he or she’.
The recoding procedure reads as follows: ‘0–6’ into ‘0’, ‘7–9’ into ‘1’, and '10–16’ into ‘2’. See Pöllmann (2008b) for information on the respective survey questions and a discussion of univariate results.
Appendix 3 provides information on the proportionate distributions of all four control variables. See Pöllmann (2008b) for information on the respective survey question, recoding procedures, and a discussion of univariate results.
It is, however, possible – and indeed desirable – to replicate (parts of) the present research which would subsequently allow for comparisons between different periods of time.
To further evaluate the quality of the model, a series of bivariate correlation analyses for all possible combinations of explanatory variables were conducted. None of the respective coefficients (Spearman's rho) was large enough to imply a problem of multicollinearity or any serious confounding of effects. See Pöllmann (2008b) for further details.
See Long and Freese (2003) for information on the computation of ideal types using Stata®.
In future research, it would be interesting to learn more about what (head) teachers mean when thinking of Europe's cultural history, with a view on possible implications for teaching and learning.
Acknowledgements
I am grateful to the 281 Berlin and London head teachers who participated in my research. I would also like to thank the ESRC for its generous support (PTA-030-2004-00357). A comprehensive list of acknowledgements can be found in Pöllman (2008b).
Appendix
. | Berlin . | London . |
---|---|---|
Intercultural capital | ||
High intercultural capital | 16.5% | 29.2% |
Medium intercultural capital | 36.7% | 33.3% |
Low intercultural capital | 46.8% | 37.5% |
Valid observations | 158 | 120 |
Countries considered as part of Europe | ||
All seven countries | 23.9% | 25.4% |
Five or six countries | 66.0% | 50.8% |
Four or fewer countries | 10.1% | 23.8% |
Valid observations | 159 | 122 |
Perceived salience of Europe's cultural history | ||
High importance | 83.0% | 77.3% |
Medium importance | 9.4% | 15.1% |
Low importance | 7.5% | 7.6% |
Valid observations | 159 | 119 |
Perceived value of EU policies | ||
Positive | 58.5% | 50.0% |
Neither negative nor positive | 32.7% | 36.9% |
Negative | 8.8% | 13.1% |
Valid observations | 159 | 122 |
. | Berlin . | London . |
---|---|---|
Intercultural capital | ||
High intercultural capital | 16.5% | 29.2% |
Medium intercultural capital | 36.7% | 33.3% |
Low intercultural capital | 46.8% | 37.5% |
Valid observations | 158 | 120 |
Countries considered as part of Europe | ||
All seven countries | 23.9% | 25.4% |
Five or six countries | 66.0% | 50.8% |
Four or fewer countries | 10.1% | 23.8% |
Valid observations | 159 | 122 |
Perceived salience of Europe's cultural history | ||
High importance | 83.0% | 77.3% |
Medium importance | 9.4% | 15.1% |
Low importance | 7.5% | 7.6% |
Valid observations | 159 | 119 |
Perceived value of EU policies | ||
Positive | 58.5% | 50.0% |
Neither negative nor positive | 32.7% | 36.9% |
Negative | 8.8% | 13.1% |
Valid observations | 159 | 122 |
Appendix
. | Berlin . | London . |
---|---|---|
Spoken foreign languages | ||
Two or more | 43.4% | 29.5% |
One | 48.4% | 45.1% |
None | 8.2% | 25.4% |
Valid observations | 159 | 122 |
Intercultural friendships | ||
Yes, several | 21.4% | 53.7% |
Yes, a few | 46.5% | 40.5% |
No, none at all | 32.1% | 5.8% |
Valid observations | 159 | 121 |
Experience of living abroad | ||
Five years or longer | 3.8% | 12.4% |
One to four years | 10.1% | 19.8% |
Less than one year | 22.8% | 12.4% |
None | 63.3% | 55.4% |
Valid observations | 158 | 121 |
. | Berlin . | London . |
---|---|---|
Spoken foreign languages | ||
Two or more | 43.4% | 29.5% |
One | 48.4% | 45.1% |
None | 8.2% | 25.4% |
Valid observations | 159 | 122 |
Intercultural friendships | ||
Yes, several | 21.4% | 53.7% |
Yes, a few | 46.5% | 40.5% |
No, none at all | 32.1% | 5.8% |
Valid observations | 159 | 121 |
Experience of living abroad | ||
Five years or longer | 3.8% | 12.4% |
One to four years | 10.1% | 19.8% |
Less than one year | 22.8% | 12.4% |
None | 63.3% | 55.4% |
Valid observations | 158 | 121 |
Appendix
. | Berlin . | London . |
---|---|---|
National attachment | ||
Very attached | 28.9% | 50.4% |
Fairly attached | 56.0% | 42.1% |
Not attached | 15.1% | 7.4% |
Valid observations | 159 | 121 |
Political alignment | ||
Far left | 8.8% | 10.7% |
Left | 55.3% | 54.9% |
Centre | 24.5% | 21.3% |
Right | 11.3% | 13.1% |
Valid observations | 159 | 122 |
Gender | ||
Female | 28.9% | 50.8% |
Male | 71.1% | 49.2% |
Valid observations | 159 | 122 |
Age | ||
56 to 65 | 41.8% | 28.7% |
47 to 55 | 46.2% | 54.9% |
37 to 46 | 12.0% | 16.4% |
Valid observations | 158 | 122 |
. | Berlin . | London . |
---|---|---|
National attachment | ||
Very attached | 28.9% | 50.4% |
Fairly attached | 56.0% | 42.1% |
Not attached | 15.1% | 7.4% |
Valid observations | 159 | 121 |
Political alignment | ||
Far left | 8.8% | 10.7% |
Left | 55.3% | 54.9% |
Centre | 24.5% | 21.3% |
Right | 11.3% | 13.1% |
Valid observations | 159 | 122 |
Gender | ||
Female | 28.9% | 50.8% |
Male | 71.1% | 49.2% |
Valid observations | 159 | 122 |
Age | ||
56 to 65 | 41.8% | 28.7% |
47 to 55 | 46.2% | 54.9% |
37 to 46 | 12.0% | 16.4% |
Valid observations | 158 | 122 |
References
Andreas Pöllmann is Postdoctoral Fellow at the National Autonomous University of Mexico (Instituto de Investigaciones Sociales). His research focuses on links between formal education and (supra) national forms of attachment; the relationship between intercultural capital and feelings of belonging; sociological perspectives on European integration; globalisation and migration; citizenship debates; advanced methods of quantitative data analysis; and comparative survey design.